#asahi-dev · 2026-01-05 14:51
jannauasahi driver down to "18 files changed, 1384 insertions(+)". probably upper bound of what I can submit right now. there is known dead code#asahi-dev · 2026-01-05 14:40
sveninitializing clusters is still inside m1n1, isn't it?#asahi-dev · 2026-01-05 14:37
fl0_id_at the point where the kernel has been uncompressed, and then says "initializing clusters" - is the device tree also important at that stage? like just wondering if it can be only chickens or a fault in the device tree could also crash it here#asahi-dev · 2026-01-05 13:41
svenit's going to be great!#asahi-dev · 2026-01-05 13:41
svenyup, and if I enable that power rail before the PHY is brought up the SoC just resets immediately#asahi-dev · 2026-01-05 13:41
mazfrom what you describe, the NHI and IOMMU are sharing a power rail (or clocks), gated by the PHY. describing that in DT is going to be fun. not.#asahi-dev · 2026-01-05 13:39
sventhanks, will take a look#asahi-dev · 2026-01-05 13:39
mazlook at the google pixel stuff (not sure it is merged yet). they definitely had some weird integration with one iommu per sswrap.#asahi-dev · 2026-01-05 13:39
sventhe weird special case here is that any access to the iommu MMIO before the NHI is powered on (which can only be done after the PHY is switched to tbt) results in an SError#asahi-dev · 2026-01-05 13:38
sveni haven't found anything in linux but maybe I just missed it#asahi-dev · 2026-01-05 13:37
mazI wonder if other SoCs that use an iommu per subsystem have similar constructs. I definitely remember something along these lines at $JOB-1...#asahi-dev · 2026-01-05 13:35
sveni'll have to ask robin and the dt guys how to properly handle this anyway#asahi-dev · 2026-01-05 13:35
sventhis isn't even the iommu for external traffic :/#asahi-dev · 2026-01-05 13:34
mazI guess that's an insurance policy. no external traffic before the IOMMU can filter things. but yeah, pretty annoying.#asahi-dev · 2026-01-05 13:33
sven*on#asahi-dev · 2026-01-05 13:33
sventl;dr: the iommu is effectively powered off until the nhi itself it brought up. and the nhi kinda depends in that iommu#asahi-dev · 2026-01-05 13:33
svenyeah... this is already annoying enough with the weird way it uses its iommu#asahi-dev · 2026-01-05 13:32
mazwell, you need to start somewhere...#asahi-dev · 2026-01-05 13:32
svenit doesn't do any pcie tunneling quite yet ofc#asahi-dev · 2026-01-05 13:32
mazsven: whooho!#asahi-dev · 2026-01-05 13:29
svenphew. finally fixed that 5 seconds reboot and the tbt nhi now shows up in boltctl again \o/#asahi-dev · 2026-01-05 13:11
svengit commit --amend --author="..." is what I usually use if I have to adjust the author to someone else. b4 will do the correct thing then#asahi-dev · 2026-01-05 13:09
Eighth_Doctorthe last from wins#asahi-dev · 2026-01-05 13:09
Eighth_Doctoryou need to override from for emails in the body#asahi-dev · 2026-01-05 12:58
integralpiloti think it's fine though, it's late here so i have to log off now#asahi-dev · 2026-01-05 12:58
integralpilotoh no git defaulted to me in From, i thought it changed it. b4 doens't use outgoing folder?#asahi-dev · 2026-01-05 12:56
integralpilotdone! :)#asahi-dev · 2026-01-05 12:56
integralpiloti've checked and i think it's alright, i'm going to send it now#asahi-dev · 2026-01-05 12:27
integralpilotmy version: https://github.com/IntegralPilot/linux-next/blob/smc-power/drivers/power/supply/macsmc-power.c#asahi-dev · 2026-01-05 12:27
integralpilotoriginal: https://github.com/AsahiLinux/linux/blob/df583f4d7a597821e124f2738eab95530ac49bb3/drivers/power/supply/macsmc-power.c#asahi-dev · 2026-01-05 12:26
integralpilotpushed to https://github.com/IntegralPilot/linux-next, smc-power branch#asahi-dev · 2026-01-05 11:48
integralpilotwill do, thank you#asahi-dev · 2026-01-05 11:47
svencan you push the branch somewhere? it's probably fine#asahi-dev · 2026-01-05 11:42
integralpilotim not sure if it's considered significant changes, though.#asahi-dev · 2026-01-05 11:24
svenas long as you don't change the from: line he's fine with his sob followed by co-dev-by. if you change the from: line or do significant changes i'd reach out before#asahi-dev · 2026-01-05 11:22
Eighth_Doctorwhile that's true, I really wouldn't recommend doing it without at least emailing him first and letting him know#asahi-dev · 2026-01-05 11:20
sven*add#asahi-dev · 2026-01-05 11:20
sventake a look at at what sob means, https://docs.kernel.org/process/submitting-patches.html#developer-s-certificate-of-origin-1-1. effectively it's enough if marcan's original version had the sob and you hadd your co-dev-by and sob afterwards#asahi-dev · 2026-01-05 11:20
integralpilotthank you!#asahi-dev · 2026-01-05 11:20
mazif there's a Co-Dev tag, you must have a SoB.#asahi-dev · 2026-01-05 11:20
jannauyou have to, it's enough that the initial version had the Signed-off-by:#asahi-dev · 2026-01-05 11:18
integralpilotdo i add Signed-off-by marcan too or no as he hasn't seen and approved the final version#asahi-dev · 2026-01-05 11:15
mazcheckpatch is not meant to be an absolute oracle. it is mostly advisory.#asahi-dev · 2026-01-05 11:15
JamesCalligeros[m]‘I’ != ‘i’ etc#asahi-dev · 2026-01-05 11:14
JamesCalligeros[m]Yeah its a fourcc it needs to be exact#asahi-dev · 2026-01-05 11:14
svenyeah, it's not perfect#asahi-dev · 2026-01-05 11:14
integralpilotthere's some false positives where it things a string literal in the SMC key macro is a subtraction (it's AC-i) and it wants spaces is that okay to ignore?#asahi-dev · 2026-01-05 11:14
integralpiloti've been using strict#asahi-dev · 2026-01-05 11:13
JamesCalligeros[m]some maintainers consider it useless without --strict and will get annoyed if you don't use it#asahi-dev · 2026-01-05 11:12
JamesCalligeros[m]I recommend running checkpatch with --strict too